Comprehensive analysis of patients with cancer with a positive CHEK2 test: Are we getting closer to a genotype/phenotype correlation?

Abstract

10588 Background: The CHEK2 gene encodes a protein kinase that acts as a tumor suppressor with an important role in DNA damage repair. This gene is included in most hereditary breast, ovarian and prostate cancer panels, but the clinical actionability of CHEK2 pathogenic/likely pathogenic variants (PVs) may be challenging. Indeed, decisions concerning clinical surveillance and/or risk reducing procedures as well as screening for multiple associated cancer risks are not consensual. Our objective was to identify a potential genotype/phenotype association regarding CHEK2 PV identified in our population. Methods: Study group included all patients (pts) who tested positive for CHEK2 PVs between July 2000 and June 2022; control group included pts without PVs after multigene panel testing (wt) between January 2020 and June 2022. PVs were categorized as: truncating (includes nonsense, frameshift, splice site variants, gross deletions and duplications) and not truncating (missense and in-frame deletions). Clinical characteristics from both groups included gender, personal and family cancer history and tumor characteristics. Results: From 8604 pts tested, we identified 101 CHEK2 and 1506 wt pts. CHEK2 pts were mainly women (95; 92%) and Portuguese (94; 93.1%). All non-Portuguese pts had Eastern Europe ancestry. CHEK2 pts were either heterozygous (94;93%), homozygous (3; 3%) or had digenic inheritance (4; 6%: both a CHEK2 and a BRCA1/2 PV). Compared to wt pts, heterozygous CHEK2 pts were younger at 1st cancer diagnosis (44.5 yrs vs 47 yrs) and had more multiple primary cancers (OR: 1.8; p: 0.02). CHEK2 breast cancer was mostly but not exclusively estrogen receptor positive (69/78; 88,4%) and Her2 negative (54/69; 78%). Two BC cases had triple negative phenotype. The most frequent PVs detected were c.349A > G (25 pts), c.1100del (24 pts) and c.319+2T > A (20 pts). Considering heterozygous PVs, 57 were truncating and 37 non-truncating, with c. 1100del and c.349A > G being the most frequent, in each group, respectively. For truncating versus non-truncating PVs there was no difference in phenotype regarding breast (OR: 0.69; p: 0.15), CRC (OR: 0.94; p: 0.91), gastric (OR: 0.76; p: 0.65), thyroid (OR: 1.3; p: 0,69) and kidney cancer (OR: 0.36; p: 0.27). Further stratification of PVs according to cancer type revealed a correlation of CRC with c.1100del (OR: 4.57; p: 0.0025) and c.349A > G (OR: 5.49; p: 0.0003), while gastric, kidney and thyroid cancers correlated with c.349A > G (OR: 5.8; p: 0.0069), c.1100del (OR: 26.86; p: 0.02) and c.319+2T > A (OR: 3.31; p: 0.05) PVs, respectively. Conclusions: This analysis suggests a potential genotype-phenotype correlation regarding CHEK2 pathogenic variants, regardless of its truncating potential. The study of clinical outcomes will further contribute to determine how this information may influence the clinical and risk management of CHEK2 pts and their families.